Cytosine base editors induce prevalent unwanted out-of-protospacer editing and target-strand editing [Targeted amplicon sequencing]

The Detect-seq provides an unbiased method for genome-wide identification of CBE induced off-targets inside of cells. Purpose: Cytosine base editors (CBEs) have the potential to correct human pathogenic point mutations. However, its genome-wide specificity remains poorly understood, precluding its therapeutic applications. Unbiased tools are urgently needed to comprehensively evaluate CBE off-targets at the genome-wide level. Methods: The genomic DNA from CBE transfected cells are processed with fragmentation, endogenous d5fC protection, damage repair, biotin-dUTP and 5fdCTP labeling, and pull-down steps. After the sequencing library preparation, the FASTQ files are produced by the Illumina HiSeq XTen platform. To guarantee the reproducibility, each case of the experiment is present with two biological replicates. Results: Through Detect-seq, we comprehensively profiled the genome-wide off-targets of CBE for several representative sgRNAs, and observed both weak, “random-like” off-targets as well as strong Cas9-dependent off-targets in different human cell lines. Cas9-dependent off-targets can be prevalent for several sgRNAs, and demonstrate a divergent degrees of sequence similarity to the sgRNA. These CBE-induced off-target sites differ greatly from those caused by Cas9 nuclease alone, suggesting inherently different properties among these genome editing tools. Targeted amplicon sequencing further corroborated the novel off-target sites by Detect-seq; quite a few sites showed editing ratio that is on the same order of magnitude or even comparable to the on-target sites. Detect-seq also revealed two unexpected types of Cas9-dependent off-targets, which are out-of-protospacer editing and target-strand editing. These novel off-targets were likely caused by an unstable structure at the PAM-distal side of the Cas9-sgRNA-DNA complex. Based upon such understanding to the off-target editing, we engineered several CBE variants bearing mutations in APOBEC1 and obtained improved CBEs with significantly reduced off-target editing activities. Overall design: Detect-seq utilized a dU-mapping strategy to profile on-target and off-target editing events by CBE at the whole-genome level. We applied Detect-seq to off-target evaluation in HEK293T and MCF7 cells for BE4max with several frequently used sgRNAs: “VEGFA site 2”, “HEK293 site 4”, “EMX1”, and “RNF2”.

Detect-seq技术提供了一种可无偏倚鉴定细胞内胞嘧啶碱基编辑器（CBE）诱导脱靶位点的全基因组分析方法。

研究背景：胞嘧啶碱基编辑器（CBE）具备校正人类致病性点突变的潜力，但其全基因组特异性仍未被充分阐明，这一局限阻碍了其临床治疗应用。因此，亟需开发无偏倚工具以在全基因组层面全面评估CBE的脱靶效应。

实验方法：从转染CBE的细胞中提取基因组DNA，依次进行片段化、内源性d5fC保护、损伤修复、生物素-dUTP与5fdCTP标记、亲和富集等步骤。完成测序文库构建后，通过Illumina HiSeq XTen平台生成FASTQ文件。为保障实验可重复性，每组实验均设置两次生物学重复。

实验结果：通过Detect-seq技术，我们对数种代表性单向导RNA（single guide RNA, sgRNA）介导的CBE全基因组脱靶位点进行了全面图谱分析，并在不同人类细胞系中观察到两类脱靶效应：一类为弱的"类随机"脱靶效应，另一类为依赖Cas9的强脱靶效应。依赖Cas9的脱靶效应在数种sgRNA中均较为普遍，且与sgRNA的序列相似性程度存在差异。CBE诱导的脱靶位点与仅由Cas9核酸酶介导的脱靶位点存在显著差异，表明这两类基因组编辑工具具有本质不同的特性。靶向扩增子测序进一步验证了Detect-seq所发现的新型脱靶位点：其中相当一部分位点的编辑效率与靶位点处于同一数量级，甚至可与靶位点相媲美。此外，Detect-seq还揭示了两种此前未被报道的依赖Cas9的脱靶类型：原间隔序列外侧编辑与靶链编辑。这类新型脱靶效应可能由Cas9-sgRNA-DNA复合物的PAM远端侧结构不稳定所导致。基于对脱靶编辑机制的这一认知，我们对载脂蛋白B mRNA编辑酶催化多肽1（APOBEC1）区域引入突变，构建了数种CBE变体，获得了脱靶编辑活性显著降低的优化型CBE。

实验整体设计：Detect-seq采用dU定位策略，在全基因组层面分析CBE介导的靶内与脱靶编辑事件。本研究针对数种常用sgRNA（包括"VEGFA位点2"、"HEK293位点4"、"EMX1"及"RNF2"），将Detect-seq应用于HEK293T与MCF7细胞中BE4max的脱靶效应评估。